Conventional glass batching processes are illustrated diagrammatically as FIG. 1. Typical glass batching involves transferring raw materials directly from storage silos into a weigh hopper, weighing the raw materials according to a weight percent (wt %) batch recipe, adding a specified amount of cullet, and mixing the raw batch and the cullet in a large scale mixer. In some cases, the mixer itself functions as a final check-scale for the batch recipe. From the mixer, the mixed batch materials are transferred to one or more hoppers positioned adjacent the end of a glass furnace (melter) where the mixed batch is introduced into the melting tank. Similar batching techniques are nearly universally employed in various glass producing industrial settings, including container glass, fiber-glass, and float glass manufacturing facilities.
After the mixed batch is added to the furnace (melter), uncontrolled reactions are allowed to occur in the melter at various elevated temperatures, both among batch raw material components and between the batch raw material components and resident melt, until a substantially homogenous melt is eventually achieved. The time required for sufficient melting, homogenization and fining defines the total residence time, or the time that the melt resides within the melter tank before being formed into the desired glass product.
FIG. 2 is a schematic illustration showing the reaction paths that the raw material batch components typically follow when reacting with each other and with the melt already present in the furnace, and FIG. 3 is a schematic illustration showing the conventionally uncontrolled melting stages as the newly added batch melts. That is, once the batch is introduced to the furnace, several reactions take place that almost immediately segregate the batch. In float glass production, for example, sodium carbonate (Na2CO3), calcium carbonate (CaCO3), sodium sulfate (Na2SO4) and quartz (SiO2) are the most commonly used major raw materials. When water has not been added in an effort to reduce batch segregation in the storage hopper, the first reaction is usually the formation of a eutectic liquid by the reaction of Na2CO3 and CaCO3 at a temperature of around 785° C.
As shown in FIG. 2, Na2CO3 and CaCO3 react along reaction Path 1, creating a low viscosity eutectic liquid with a quantity of un-reacted CaCO3. This low viscosity eutectic liquid reacts with residual CaCO3 and quartz along reaction Path 2 to eventually achieve the overall composition of the glass dictated by the batch recipe. An example of a typical float glass composition is approximately 73.5 wt. % SiO2, 12.3 wt. % CaO, and 14.2 wt. % Na2O.
Similar reactions are observed between Na2CO3, CaCO3, and Na2SO4. In this case, the eutectic liquid is composed of molten salts having a very low viscosity. That is, the eutectic liquid flows easily, and exhibits flow properties similar to those exhibited by water, which has a viscosity in a range of 1 to 4 mPa·s, or 1 to 4 centipoise. The eutectic liquid reacts with the quartz to eventually provide a homogeneous glass of the desired composition. The formation of this eutectic liquid, however, can increase the tendency for batch segregation within the melter, as the eutectic liquid has very low viscosity and easily flows downhill in the melter to pool at the bottom, leaving an island of high melting point quartz sand. This effectively reverses the efforts of mixing prior to introducing the batch into the melter.
Similar reactions occur in container glass compositions, and in the case of borosilicate fiber-glass production, borate raw materials contribute to similar problems in the initial stages of melting. This segregation process leads to the formation of large-scale domains, or agglomerates, of nearly pure silica that then require excessively long residence times for dissolution into the surrounding liquid melt. This initial segregation then requires re-homogenization within the glass tank prior to forming a homogeneous melt.
Direct evidence of “de-mixing” can be seen in a glass tank during the melting process. Agglomerations (scaled on the order of centimeters in length) of batch raw materials, commonly referred to in the industry as batch logs or turtles, can be seen in various states of melting in the glass tank. Moreover, the phenomenon of large-scale batch segregation in the melter tank is commonly seen in finished glass in the form of defects such as stones, which are mostly composed of undissolved quartz; seeds, which are bubbles that are not liberated from the melt during fining; and cord lines, which are optical distortions caused by local differences in composition. These defects are direct evidence of off-composition glass due to batch de-mixing or incomplete re-mixing that decrease the overall energy and material efficiency of the melter and reduce the quality of the final product. Industrial observations are further supported by technical publications, which also recognize that batch segregation is commonly observed in commercial production. Despite the fact that batch segregation in the (glass tank and the potential defects that can result therefrom) are recognized in the industry, and despite a long felt need to reduce this undesirable behavior and improve melting efficiency and overall quality, the glass industry has not yet successfully addressed these issues with a viable commercial solution.
As mentioned above, material efficiency in glass making is related to reducing losses due to defects such as stones, seeds, and cord lines. Stones are silica particles or agglomerates that have not fully reacted with the melt. The appearance of this type of defect can be decreased by reducing segregation of refractory silica from flux materials. Seeds, which are bubbles that result from incomplete fining, can be reduced by maximizing the timing of the evolution of volatiles early in the melting process and by reducing air trapped in pore spaces. While cullet from some defective glass can be recycled through the process (though glass with stones cannot be recycled), it is more efficient to reduce the amount of in-house cullet from defective glass.
In large scale commercial glass production (e.g., flat glass, container glass, and fiber-glass) where the melting tank volumes are considerably greater (accommodating volumes on the order of tons of molten glass), in situ melt mixing is accomplished by convection currents within the tank and by the movement of evolved gases from decomposition of raw materials. While some mixing and fining is required to remove gaseous bubbles, the expensive and energy intensive processes to improve the mixing of the molten batch can also be attributed to large scale segregation of batch materials.
Considering that physical mixing is but one factor, the efficiency of the melting process is therefore directly related to diffusion or reactions at the quartz-liquid interface. Quartz dissolution is limited by the initial reaction of quartz with the low viscosity eutectic liquid. As melting progresses, the quartz interacts with a liquid that is steadily increasing in silica content and subsequently, viscosity. Therefore, high temperatures are needed within the melting tank to ensure reasonable diffusion rates and reasonable homogeneity. As mentioned above, the residence time of the material in a tank is determined by the time it takes for the batch materials to completely melt and for the resulting liquid to homogenize. In a continuous production situation, the mass of molten glass in the furnace is held constant, and commercially, the minimum mean residence time is of the order of 24 hours of production for container furnaces and 72 hours for flat glass furnaces with roughly half of this time devoted to melting and the other half devoted to fining.
One attempt to improve the batch melting process involved reducing the addition of carbonate and quartz in the raw (unmixed) form. Experiments were conducted using synthetic diopside (CaO MgO 2SiO2) instead of a mixture of CaCO3, MgCO3 and quartz. The results showed that the time required to completely dissolve the original batch (i.e., the batch free time) was reduced depending on temperature, and there was also a reduction in fining time. These improvements were attributed to a reduction in the amount of quartz requiring dissolution into the melt.
It is also conventional wisdom that melting (more accurately, bubble removal) is promoted by keeping the viscosity low. As described above, however, the uncontrolled production of low viscosity liquids during the melting process contributes to undesirable batch segregation. Although a melt that fosters lower viscosities overall may improve quartz dissolution and diffusion rates during melting, these benefits can only be achieved after the highest melting point batch components are sufficiently melted and any batch agglomerates are fully reacted in the melt. Thus, in order to improve melting efficiency and reduce the above-described problems associated with de-mixing and segregation, substantial improvements with respect to controlling the glass batch melting behavior are desired.
Another drawback with conventional glass making technology arises from the amount of energy required to maintain a continuous glass melting operation, and the environmental impact of the use of the fossil fuel necessary to provide this energy. Fuel can constitute 25-30% of the cost of manufacturing float glass. The volatility of fuel prices can, of course, at times increase this proportion without warning. Nationwide, the U.S. glass industry uses in excess of 250 trillion BTU annually to produce approximately 21 million tons of glass products; approximately 80% of this energy is supplied by natural gas. Melting one ton of glass theoretically requires only about 2.2 million BTU, but in practice the fuel requirement ranges from about 4.7 to about 6.9 million BTU per ton due to losses and inefficiencies. Because 80% or more of the overall energy used in the manufacture of container glass, fiber-glass, and flat glass is needed for the melting and fining operations, an energy reduction in glass manufacturing through more efficient melting would be desirable. For example, if a flat glass plant producing 400 tons/day of flat glass runs 365 days/year, even the most efficient natural gas-fired plant (4.7 million BTU/ton) consumes approximately 686 billion BTU/year, or 686 million cubic feet of natural gas.
Pollution prevention and reduction of the considerable costs associated with regulatory compliance, as well as improving the overall energy and material efficiency, are critical for reducing the negative environmental impact of glass manufacturing and for making glass manufacturing more economically competitive. For example, a typical float glass plant must spend an average of $2 million dollars for new environmental control systems and about 2.5% of total manufacturing costs on compliance. Thus, a reduction in 10% of the natural gas use in a typical float plant would result in a savings of approximately $285,000 per year in natural gas (conservatively assuming $5/MMBtu). Moreover, reductions in compliance costs associated with additional chemical treatments and operational implementations aimed at reducing pollutant emissions from combustion reactions could also be realized in conjunction with a reduction in the amount of fuel consumed.
Air pollutants emitted from glass industry include:
1) Nitrogen oxides (NOx)2) Sulfur oxides (SOx)3) Carbon monoxide (CO)4) Carbon Dioxide (CO2)
Fossil fuels used for combustion are typically the source of NOx and some COx. The decomposition of carbonate and sulfate raw materials likewise contribute COx and SOx emissions, respectively. Reduction of the residence time of the melt, however, was reduced to the amount of fuel burned per unit of glass produced and improved energy efficiency. This would also foster reduced amounts of emissions such as NOx and fuel-derived CO2 and CO per unit of glass produced.
Residence time is related to the time required to fully melt all of the batch components, and is particularly dependent upon the amount of high-melting point batch components (e.g., silica) in the batch recipe. Although it would be desirable to eliminate free quartz as a raw material additive due to its slow reactivity and high melting point, quartz remains an abundant and economical source of silica, which is a major component of many commercial glass systems.
Boron-containing glass melts, such as E-glass, can present some particular challenges in the glass melting process. A typical E-glass system is comprised of four main chemical components: B2O3, SiO2, Al2O3, and CaO. Granular boric acid is one of the major sources for B2O3 used in the making of E-glass, especially where low sodium content is required in the glass composition. This low sodium content is particularly important in the reinforcement fiber glass industry and in the production of glass for LCD panel displays.
When calcium (e.g. as calcium carbonate) and boron (e.g. as boric acid, boric oxide, or a borate salt), are included in glass formulations low viscosity melts can form at relatively low temperatures in the glass melting process. According to traditional practice, when batch is introduced to the melting furnace several reactions take place that almost immediately segregate the batch. Boric acid (H3BO3) melts and forms a low viscosity liquid at about 170° C. Both the apparent melt temperature and the viscosity can be substantially reduced by the presence of ambient moisture. Many alkali metal borates exhibit similar incongruent melt behavior, forming a liquid phase at temperatures well below the melt temperature of the final glass composition. Upon further heating, boric acid loses water and forms meta-boric acid (HBO2). With further heating, it forms an anhydrous boric oxide (B2O3) melt below 500° C. Calcium carbonate releases carbon dioxide forming calcium oxide, which also melts at a relatively low temperature and can flow away from the rest of the batch.
Boric oxide has dual roles in the glass making process; it behaves as a flux during initial melt formation and as a glass former in the structure formation phase. Liquid B2O3 enhances dissolution of refractory batch constituents in boron-containing glass formulations such as E-glass. However, the appearance of a low viscosity borate melt in the initial stages of the glass making process can cause segregation in the glass tank during melting, thus removing the fluxing agent away from the refractory batch constituents, and limiting its contribution to the dissolution of the remaining batch materials. This process of segregation leads to large-scale zones or agglomerates of nearly pure silica, and potentially, agglomerates of other raw materials, such as calcium carbonate, wollastonite, alumino-silicates (such as kyanite or meta-kaolin), that can require excessively long residence times in large glass melt tanks for dissolution into the surrounding liquid melt, such as on the order of nine hours at 1300° C., higher temperatures, and/or physical mixing aids. Each of these remedial steps results in higher energy use and more time to successfully produce a homogeneous glass melt.
In some cases, melt efficiency suffers due to the formation of refractory crystalline material in the melt. Other problems caused by the presence of liquid B2O3 in the melting furnace relate to volatile losses of boric oxide. If such volatile losses are excessive, such as when long residence times are required in the melting furnace, it can result in environmental concerns or increased pollution abatement equipment costs, as well as added raw material costs.
Thus, it would be desirable to provide a method for controlling the melting behavior (i.e., reaction paths) of glass batch components within a resident melt to improve melting efficiency, such that the improved melting efficiency enables a decrease in energy usage, reduces the need for chemical fining agents that contribute to air pollutants and raw material cost, decreases pollution while ultimately producing higher quality, lower cost glass products and reduces the occurrence of batch de-mixing and segregation in early melting stages. The novel technology discussed herein addressed this need.